What type of study is this?

August 17, 2025Open Access

Multimodal Deep Learning Framework for Alzheimer’s Disease Classification Using MRI Scans and Clinical Records

Key Points

AlzFusionNet achieved a classification accuracy of 96.8% for Alzheimer's disease stages.
The model outperformed single-modality approaches by integrating MRI features and clinical data.
This analysis employed PCA for dimensionality reduction and efficientnet-b7 for MRI feature extraction.
The findings highlight the effectiveness of multimodal integration for improved Alzheimer’s disease classification.

Abstract

Alzheimer's disease (AD) is a neurodegenerative disorder that strikes millions of people globally. Classification that is both accurate and early-stage is key to successful intervention and treatment planning. In this paper, AlzFusionNet, a multimodal deep learning model fusing MRI features with clinical information for enhanced AD classification, is proposed. EfficientNet-B7 is used as the backbone of the model to extract MRI features, and PCA is used to decrease the dimensionality of clinical information. A fusion layer merges both modalities for four-stage classification: Non-Demented, Very Mild Demented, Mild Demented, and Moderate Demented. Experimental outcomes prove that AlzFusionNet outperforms single-modality models in terms of accuracy (96.8%). The results underscore the advantages of multimodal integration for AD classification.

Read Full Paperexternally

Ask AI

Helpful

Bookmark

View Full Paper